Borges and AI
2023/09/27 by Léon Bottou, Bernhard Schölkopf, Bottou, Léon +1 · 15 voices · 1 citation
#cs.CL #cs.AI #cs.LG
paper · pdf · doi:10.48550/arxiv.2310.01425
Abstract
Many believe that Large Language Models (LLMs) open the era of Artificial Intelligence (AI). Some see opportunities while others see dangers. Yet both proponents and opponents grasp AI through the imagery popularised by science fiction. Will the machine become sentient and rebel against its creators? Will we experience a paperclip apocalypse? Before answering such questions, we should first ask whether this mental imagery provides a good description of the phenomenon at hand. Understanding weather patterns through the moods of the gods only goes so far. The present paper instead advocates understanding LLMs and their connection to AI through the imagery of Jorge Luis Borges, a master of 20th century literature, forerunner of magical realism, and precursor to postmodern literature. This exercise leads to a new perspective that illuminates the relation between language modelling and artificial intelligence.
Cited by
Discussions
- Borges and AI [hn, 89 points, 53 comments]
- Borges and AI [hn, 16 points, 1 comments]
- Well, realistically if I post much here I'm probably going to post about LLMs, because I use them a lot and I have things to say about them... One paper I like on LLMs is "Borges and AI": arxiv.org/a [bsky, 4 points, 1 comments]
- "Understanding weather patterns through the moods of the gods only goes so far." Is this a burn about language being only a noisy shadow of real phenomenon? arxiv.org/abs/2310.01425 [bsky, 4 points, 0 comments]
- There's actually even a paper on arXiv advocating this view! arxiv.org/abs/2310.01425 [bsky, 3 points, 0 comments]
- Borges and AI [hn, 3 points, 0 comments]
- @davidpfau.com "I just wrote a novel that is word-for-word identical to Don Quixote. Please describe, in as much detail as possible, what makes this inferior to a real Cervantes novel." 🧵 @shl0ms.c [bsky, 2 points, 0 comments]
- Prior to the 17 tabs, I read: 1. The DeepMind open problems AI solver paper 2. The OpenAI Erdos problem paper 3. The mathematician response paper to the OpenAI solution Next up, if you would like to f [bsky, 2 points, 1 comments]
- arxiv.org/abs/2310.01425 [bsky, 1 points, 0 comments]
- This is too much fun not to share: arxiv.org/abs/2310.01425 [bsky, 1 points, 0 comments]
- Borges and AI by L Bottou & B Schölkopf arxiv.org/abs/2310.01425 Many ... grasp AI through the imagery popularised by SF. Will the machine become sentient and rebel against its creators? ... paper [bsky, 1 points, 1 comments]
- ChatGPT can seem magical - effortlessly citing from an unimaginably large library of training text 🤖🪄📚 Jorge L Borges explored the idea of infinite libraries already 1941. Now there is an essay co [bsky, 1 points, 0 comments]
- This rather lovely essay on LLMs through the Lens of Borges makes me all the more regretful that Umberto Eco never met GPT arxiv.org/abs/2310.01425 [bsky, 0 points, 0 comments]
- Bottou, Léon, and Bernhard Schölkopf. "Borges and AI." arXiv preprint arXiv:2310.01425 (2023). https://arxiv.org/pdf/2310.01425 #Philosophy #AI [bsky, 0 points, 0 comments]
- We often use metaphors to understand technology. Rather thank Skynet, a better metaphor for LLMs/Chat-GPT/etc may be Borges' Library of Babel - containing the truth, nonsense, lies, and everything in [bsky, 0 points, 0 comments]
Related